Connect your AI to your Infrastructure with MCP

It’s no longer about manually building workflows. Now you can give your AI direct and secure access to your n8n infrastructure hosted on AIBuildr.

Thanks to the Model Context Protocol (MCP) and the tools developed by Mateusz Czlonkowski — n8n-mcp and n8n-skills — your AI can understand, read, and execute your workflows like an expert engineer.

What do you get with this?

• Real Execution: Your AI can execute nodes, test webhooks, and query your database via n8n.
• Total Privacy: Your credentials and sensitive data remain on your AIBuildr instance. The AI only receives the necessary context.
• Autonomous Agents: Create agents that can “see” your tools and use them to solve complex problems.

---

Tool 1: n8n-MCP Server

The bridge connecting your AI with n8n. This MCP server gives your AI access to documentation for over 1,084 n8n nodes (537 core + 547 community) and, optionally, full workflow management via the REST API.

Prerequisites

• Node.js installed on your local machine.
• Your n8n URL (e.g., https://n8n.yourdomain.aibuildr.tech).
• An n8n API Key (generate it in Settings > Public API).

Configuration per Tool

Configuration uses environment variables, not CLI flags. The key variables are:

Variable	Description
MCP_MODE	Required. Must be "stdio" to avoid JSON errors.
N8N_API_URL	Your n8n instance URL.
N8N_API_KEY	Your n8n API Key.
LOG_LEVEL	Recommended: "error" to reduce noise.
DISABLE_CONSOLE_OUTPUT	Recommended: "true".

---

1. Claude Desktop

Edit your configuration file:

• Mac: ~/Library/Application Support/Claude/claude_desktop_config.json
• Windows: %APPDATA%\Claude\claude_desktop_config.json
• Linux: ~/.config/Claude/claude_desktop_config.json

{
  "mcpServers": {
    "n8n-mcp": {
      "command": "npx",
      "args": ["n8n-mcp"],
      "env": {
        "MCP_MODE": "stdio",
        "LOG_LEVEL": "error",
        "DISABLE_CONSOLE_OUTPUT": "true",
        "N8N_API_URL": "https://n8n.yourdomain.aibuildr.tech",
        "N8N_API_KEY": "your-api-key-here"
      }
    }
  }
}

Restart Claude Desktop. You will see a plug 🔌 icon indicating the connection.

---

2. Cursor (AI IDE)

Create a .cursor/mcp.json file in your project root:

{
  "mcpServers": {
    "n8n-mcp": {
      "command": "npx",
      "args": ["n8n-mcp"],
      "env": {
        "MCP_MODE": "stdio",
        "LOG_LEVEL": "error",
        "DISABLE_CONSOLE_OUTPUT": "true",
        "N8N_API_URL": "https://n8n.yourdomain.aibuildr.tech",
        "N8N_API_KEY": "your-api-key-here"
      }
    }
  }
}

Then, go to Cursor Settings > MCP and enable the server.

---

3. Antigravity

Antigravity natively supports MCP, allowing it to “see” your n8n workflows as local tools.

Step-by-step Setup:

1. Install n8n-mcp globally in your terminal:

   npm install -g n8n-mcp

2. Open the Antigravity chat and click the three dots ... > MCP Servers > Manage MCP Servers.
3. Click View raw config to open the mcp_config.json file.
4. Add the following configuration to the mcpServers object:

{
  "mcpServers": {
    "n8n-mcp": {
      "command": "node",
      "args": [
        "<GLOBAL_PATH_TO_n8n-mcp>/dist/mcp/index.js"
      ],
      "env": {
        "MCP_MODE": "stdio",
        "LOG_LEVEL": "error",
        "DISABLE_CONSOLE_OUTPUT": "true",
        "N8N_API_URL": "https://n8n.yourdomain.aibuildr.tech",
        "N8N_API_KEY": "your-api-key-here"
      }
    }
  }
}

5. Go back to Manage MCP Servers and click Refresh. The server will appear with a green dot, indicating it is active.

---

Can I install a skill.md in an n8n agent?

This is one of the most common questions when integrating AI with n8n. The short answer is YES, but the way to do it depends on how you are building your agent:

1. Using the n8n AI Agent Node

If you use the native AI Agent node, you don’t “install” an .md file physically. Instead, you copy the content of the skill.md and paste it into the System Prompt field. This gives the AI specific instructions on how to behave and use the tools.

2. Using n8n-skills (Recommended)

The n8n-skills repository already contains structured best practices. You can ask an AI that already has access to these skills (via MCP or by installing them locally) to use them to configure your n8n nodes optimally.

3. Automation with n8n

You can create a workflow in n8n that reads .md files from a GitHub repository or local storage and injects them dynamically into your AI nodes via variables or expressions.

---

Tool 2: n8n-skills

If n8n-mcp are the hands of the AI, n8n-skills is the brain training. It is a set of 7 complementary skills so that Claude knows how to build in n8n following best practices.

The 7 Skills

#	Skill	What for?
1	Expression Syntax	Expression syntax $json, $node, $now
2	MCP Tools Expert	Correct usage of MCP tools (highest priority)
3	Workflow Patterns	5 proven architectural patterns
4	Validation Expert	Interpret validation errors and fix them
5	Node Configuration	Complex node configuration and dependencies
6	Code JavaScript	JavaScript in n8n Code nodes
7	Code Python	Python in Code nodes (for the 5% of cases)

Installation

For Claude Code (recommended):

# Install directly as a plugin
/plugin install czlonkowski/n8n-skills

Manual (any environment):

git clone https://github.com/czlonkowski/n8n-skills.git
cp -r n8n-skills/skills/* ~/.claude/skills/

---

Your Infrastructure on AIBuildr

For all this to work with minimal latency and maximum security, your n8n instance must be robust. On AIBuildr, your n8n runs on optimized infrastructure, with automatic backups and dedicated resources — perfect for supporting your AI agents’ workload.

---

Troubleshooting

Error: “Unexpected token…”

Make sure to include "MCP_MODE": "stdio" in your configuration’s environment variables. Without this, debug logs interfere with the JSON-RPC protocol.

Error: “Connection refused”

Ensure your n8n instance on AIBuildr is active and publicly accessible. AIBuildr automatically configures SSL, so make sure to use https://.

Error: “Invalid API Key”

Generate a new API Key in n8n:

1. Enter your n8n panel.
2. Go to Settings > Public API.
3. Create a new key and copy it immediately.